Two months after Anthropic introduced Mythos to a limited group of users, citing concerns that the AI model could be misused if widely released, the company says it is now prepared to launch a similarly powerful system to the public.

On Tuesday, Anthropic announced Claude Fable 5, a Mythos-class model that will be made available to enterprise customers and paid subscribers.

The company said the wider rollout is enabled by new safety measures designed to restrict outputs in sensitive areas such as cybersecurity and biology.

“For us, it’s really around what we call ‘race to the top,’ being able to provide this technology in a valuable fashion, and at the same time providing the right safety guardrails so that it can do asymmetrically more benefits than harm,” Dianne Penn, Anthropic’s head of product management for research, told CNBC.

Anthropic drew attention from Wall Street and government officials in April when it unveiled Mythos, a model designed to excel at detecting security vulnerabilities in software. The company said it had no plans for a general public release and instead restricted access to a small number of firms under a cybersecurity program known as Project Glasswing.

With the launch of Claude Fable 5, Anthropic is moving closer to what it has described as its “eventual goal” of deploying Mythos-class models at scale.

The rollout also comes as investor interest in its technology continues to build ahead of a potential blockbuster IPO, which could happen as early as this year.

The company said Claude Fable 5 delivers “exceptional performance” in software engineering and knowledge work tasks.

According to a blog post, it outperforms Claude Opus 4.8, another model introduced late last month, by more than 10 per month on certain benchmarks.

Penn said Claude Fable 5 marks a “significant jump” in capability, which is why the company has introduced additional safeguards to reduce the risk of misuse.